Size, timing and source matter more than the text.
Running on the fast 2-second loop, the clipboard detector never needs to read what was pasted to know the event may matter. It evaluates metadata: how much, how structured, when, in which foreground app, and what else was happening.
Small pastes during coding are routine. A large structured paste seconds after the hardest question, with an AI tab active, is the pattern that matters.
Large paste bursts
Flags large chunks of text or code that appear suddenly during a live answer.
Question timing
Correlates paste events with interview questions, long pauses, and answer bursts.
Signal correlation
A paste after AI-tool activity, a focus switch, or remote-control evidence gets stronger review context.
Legitimate example
A candidate pastes a pre-approved boilerplate snippet during an open-book coding task. The event is visible, contextual, and not paired with other high-risk signals.
High-risk example
An AI assistant appears, the candidate pauses, focus changes, and a large structured answer is pasted into the editor seconds later.
Clipboard detection questions, answered.
Does InterviewWatch read what is on the clipboard?
Isn't pasting normal during coding?
How is a clipboard signal scored?
Catch pasted answers without reading a keystroke.
Add privacy-preserving clipboard-context signals to your live coding rounds.
Try nowRead the coding-test guide